Weekly investment commentary en us 20260727 cheaper AI new earnings questions

CAPITAL AT RISK. INVESTMENTS CAN RISE OR FALL IN VALUE. FOR PUBLIC DISTRIBUTION IN THE U.S., CANADA, LATIN AMERICA, ​ SELECT COUNTRIES IN EUROPE (SEE THE FULL DISCLAIMER), ISRAEL, SOUTH AFRICA, HONG KONG, SINGAPORE AND AUSTRALIA. ​ FOR INSTITUTIONAL, PROFESSIONAL, AND QUALIFIED INVESTORS AND CLIENTS IN OTHER PERMITTED COUNTRIES.

Weekly commentary July 27, 2026

Cheaper AI, new earnings questions Jean Boivin • U.S. earnings remain exceptionally strong. But the focus should also be on AI Head – BlackRock profit durability, not just another round of earnings beats. Investment Institute • Geopolitics drove markets last week. Renewed Middle East supply risks and U.S. tariff tensions pushed oil briefly above $100 a barrel and Treasury yields higher. Wei Li • The Federal Reserve takes center stage this week. Its rate decision, alongside Global Chief Investment U.S. GDP and PCE inflation data, could reinforce our high-for-longer rate view. Strategist – BlackRock Investment Institute This earnings season comes with unusually high expectations. Consensus expects a second consecutive quarter of more than 20% S&P 500 earnings growth, driven Carrie King largely by the AI buildout. Early results have again exceeded forecasts. Yet the Global Chief Investment bigger question is whether today’s extraordinary profit levels can be sustained as Officer, Fundamental Equities – BlackRock cheaper models reshape the economics of AI, a key theme in our Midyear Outlook. The clues are more likely to come from earnings calls than from headline results. Natalie Gill Senior Portfolio Strategist High expectations – BlackRock Investment S&P 500 12-month forward earnings, 2012-28 Institute 6.5 July 2028 6

Forward-looking estimates may not come to pass. Source: BlackRock Investment Institute with data from LSEG Datastream, July 24, 2026. Notes: The dotted line shows the 20-year linear trend on a logarithmic scale, representing the long-run compound earnings growth path. The July 2028 projection is based on the latest 12-month forward earnings estimate and Visit BlackRock Investment consensus expected earnings growth from months 12 to 24. Institute for insights on the U.S. equities do not look especially expensive on a forward price-to-earnings global economy, markets basis, but cyclically adjusted (Shiller CAPE) valuations remain historically rich. and geopolitics. The difference reflects consensus expectations that today’s extraordinary earnings growth, supported by the unprecedented speed and scale of the AI buildout, will persist rather than revert to historical norms. See the chart. Whether that assumption holds is the question. The emergence of powerful Chinese AI models, including Moonshot's Kimi K3, could put it to the test. As competition intensifies, the debate is no longer just who will build the best model, but who will capture the economic rent. We think cheaper AI changes the winners, not the investment case. Instead, it reinforces our preference for AI infrastructure over the increasingly competitive model layer.

FOR PUBLIC DISTRIBUTION IN THE U.S.,CAPITAL CANADA,ATLATIN RISK. AMERICA, AUSTRIA, INVESTMENTS CANGERMANY, FRANCE, RISE OR FALL ITALY, IN VALUE. LIECHTENSTEIN, FOR IRELAND,INSPAIN, PUBLIC DISTRIBUTION PORTUGUAL, THE U.S., BELGIUM, CANADA, LATIN UK, LUXEMBOURG, AMERICA, ​ SWITZERLAND, NETHERLANDS, NORWAY, SELECT FINLAND, COUNTRIES SWEDEN,(SEE IN EUROPE DENMARK, ISRAEL, THE FULL SOUTH AFRICA, DISCLAIMER), ISRAEL,HONG SOUTHKONG, SINGAPORE AFRICA, AND AUSTRALIA. HONG KONG, SINGAPOREFORANDINSTITUTIONAL, AUSTRALIA. ​ FORPROFESSIONAL, INSTITUTIONAL,QUALIFIED INVESTORS PROFESSIONAL, AND AND QUALIFIED QUALIFIED CLIENTS INVESTORS IN OTHER AND PERMITTED CLIENTS IN COUNTRIES.COUNTRIES. OTHER PERMITTED

The cost of AI is emerging as a key concern for companies deploying it. Gartner expects worldwide spending on AI models and platforms to reach $64 billion in 2026, up 63% from 2025. As enterprise AI bills rise, companies have a stronger incentive to contain costs through model routing and lower-cost models. Reflecting that shift, OpenRouter data on the 10 most widely used AI models show Chinese models processing roughly 23 trillion tokens a week…